The overall hypothesis is that genomes record ancestry because genomes are almost perfect copies of copies. The greater the numbers of divisions since a "start" or common ancestor, on average the greater the number of differences between genomes (a molecular clock hypothesis). Although molecular phylogeny is commonly used to reconstruct the past of species and populations, this approach has not been translated to normal or neoplastic somatic human cells. Molecular phylogeny is uniquely suited to reconstruct human somatic cell ancestry because it requires no prior experimental intervention[unreadable]genomes surreptitiously record ancestry through replication errors. More conventional cell fate markers used in experimental systems such as mice require the prior introduction of genetic markers, which is impractical for humans. Therefore, molecular phylogeny, which does not require prior experimental manipulations, is one of the few practical approaches to reconstruct the histories of human somatic cells. This Project hypothesizes that tumor genomes become polymorphic after transformation, and these variations record ancestry, or how and how fast it took for a single transformed cell to become the present day tumor population. We will apply this hypothesis to analyze human acute myelogenous leukemia specimens. The research will further develop epigenetic somatic cell molecular clocks based on DNA methylation patterns, which appear to drift fast enough to record somatic cell ancestries. These studies will complement the mouse studies of the other Projects, and will help translate the mouse studies to human diseases. Success of this Project will provide a systematic method to take any human tumor and look back in time to reconstruct its ancestry from variations in its cancer genomes.